i need to remind you that my academic background is not in politics or philosophy or screwing around and having fun. it is in mathematics. i am interested in policy, and i don't have much interest in competition, but i find the idea of approaching the election as a math problem to be fascinating. that's why i'm bothering. i've been over this elsewhere. it's not exactly a typical approach.
the support for the idea that race is predictive just isn't there. i've railed against racist media narratives. but, it may also be a little bit of laziness all around. it may be attempting to try and reuse ideas from 2008.
http://www.salon.com/2016/04/01/bernie_sanders_may_be_leading_hillary_clinton_with_black_voters_in_this_one_very_important_state/